Skip to content

Commit 0834905

Browse files
serhiy-storchakaserhiy-storchaka
authored
[3.6] bpo-13617: Reject embedded null characters in wchar* strings. (GH-2302) (#2462)
Based on patch by Victor Stinner. Add private C API function _PyUnicode_AsUnicode() which is similar to PyUnicode_AsUnicode(), but checks for null characters.. (cherry picked from commit f7eae0a)
1 parent 413c0a9 commit 0834905

22 files changed

Lines changed: 115 additions & 24 deletions

‎Include/unicodeobject.h‎

Lines changed: 7 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -752,23 +752,27 @@ PyAPI_FUNC(Py_UCS4*) PyUnicode_AsUCS4(
752752
PyAPI_FUNC(Py_UCS4*) PyUnicode_AsUCS4Copy(PyObject *unicode);
753753
#endif
754754

755+
#ifndef Py_LIMITED_API
755756
/* Return a read-only pointer to the Unicode object's internal
756757
Py_UNICODE buffer.
757758
If the wchar_t/Py_UNICODE representation is not yet available, this
758759
function will calculate it. */
759760

760-
#ifndef Py_LIMITED_API
761761
PyAPI_FUNC(Py_UNICODE *) PyUnicode_AsUnicode(
762762
PyObject *unicode /* Unicode object */
763763
);
764-
#endif
764+
765+
/* Similar to PyUnicode_AsUnicode(), but raises a ValueError if the string
766+
contains null characters. */
767+
PyAPI_FUNC(const Py_UNICODE *) _PyUnicode_AsUnicode(
768+
PyObject *unicode /* Unicode object */
769+
);
765770

766771
/* Return a read-only pointer to the Unicode object's internal
767772
Py_UNICODE buffer and save the length at size.
768773
If the wchar_t/Py_UNICODE representation is not yet available, this
769774
function will calculate it. */
770775

771-
#ifndef Py_LIMITED_API
772776
PyAPI_FUNC(Py_UNICODE *) PyUnicode_AsUnicodeAndSize(
773777
PyObject *unicode, /* Unicode object */
774778
Py_ssize_t *size /* location where to save the length */

‎Lib/ctypes/test/test_loading.py‎

Lines changed: 2 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -62,6 +62,8 @@ def test_load_library(self):
6262
windll["kernel32"].GetModuleHandleW
6363
windll.LoadLibrary("kernel32").GetModuleHandleW
6464
WinDLL("kernel32").GetModuleHandleW
65+
# embedded null character
66+
self.assertRaises(ValueError, windll.LoadLibrary, "kernel32\0")
6567

6668
@unittest.skipUnless(os.name == "nt",
6769
'test specific to Windows')

‎Lib/test/test_builtin.py‎

Lines changed: 6 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -151,6 +151,8 @@ def test_import(self):
151151
self.assertRaises(TypeError, __import__, 1, 2, 3, 4)
152152
self.assertRaises(ValueError, __import__, '')
153153
self.assertRaises(TypeError, __import__, 'sys', name='sys')
154+
# embedded null character
155+
self.assertRaises(ModuleNotFoundError, __import__, 'string\x00')
154156

155157
def test_abs(self):
156158
# int
@@ -1002,6 +1004,10 @@ def test_open(self):
10021004
self.assertEqual(fp.read(300), 'XXX'*100)
10031005
self.assertEqual(fp.read(1000), 'YYY'*100)
10041006

1007+
# embedded null bytes and characters
1008+
self.assertRaises(ValueError, open, 'a\x00b')
1009+
self.assertRaises(ValueError, open, b'a\x00b')
1010+
10051011
def test_open_default_encoding(self):
10061012
old_environ = dict(os.environ)
10071013
try:

‎Lib/test/test_curses.py‎

Lines changed: 10 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -81,7 +81,7 @@ def test_window_funcs(self):
8181
win2 = curses.newwin(15,15, 5,5)
8282

8383
for meth in [stdscr.addch, stdscr.addstr]:
84-
for args in [('a'), ('a', curses.A_BOLD),
84+
for args in [('a',), ('a', curses.A_BOLD),
8585
(4,4, 'a'), (5,5, 'a', curses.A_BOLD)]:
8686
with self.subTest(meth=meth.__qualname__, args=args):
8787
meth(*args)
@@ -194,6 +194,15 @@ def test_window_funcs(self):
194194
self.assertRaises(ValueError, stdscr.instr, -2)
195195
self.assertRaises(ValueError, stdscr.instr, 2, 3, -2)
196196

197+
def test_embedded_null_chars(self):
198+
# reject embedded null bytes and characters
199+
stdscr = self.stdscr
200+
for arg in ['a', b'a']:
201+
with self.subTest(arg=arg):
202+
self.assertRaises(ValueError, stdscr.addstr, 'a\0')
203+
self.assertRaises(ValueError, stdscr.addnstr, 'a\0', 1)
204+
self.assertRaises(ValueError, stdscr.insstr, 'a\0')
205+
self.assertRaises(ValueError, stdscr.insnstr, 'a\0', 1)
197206

198207
def test_module_funcs(self):
199208
"Test module-level functions"

‎Lib/test/test_grp.py‎

Lines changed: 2 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -50,6 +50,8 @@ def test_errors(self):
5050
self.assertRaises(TypeError, grp.getgrgid)
5151
self.assertRaises(TypeError, grp.getgrnam)
5252
self.assertRaises(TypeError, grp.getgrall, 42)
53+
# embedded null character
54+
self.assertRaises(ValueError, grp.getgrnam, 'a\x00b')
5355

5456
# try to get some errors
5557
bynames = {}

‎Lib/test/test_imp.py‎

Lines changed: 4 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -314,6 +314,10 @@ def test_multiple_calls_to_get_data(self):
314314
loader.get_data(imp.__file__) # File should be closed
315315
loader.get_data(imp.__file__) # Will need to create a newly opened file
316316

317+
def test_load_source(self):
318+
with self.assertRaisesRegex(ValueError, 'embedded null'):
319+
imp.load_source(__name__, __file__ + "\0")
320+
317321

318322
class ReloadTests(unittest.TestCase):
319323

‎Lib/test/test_locale.py‎

Lines changed: 5 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -339,9 +339,14 @@ def test_strcoll(self):
339339
self.assertLess(locale.strcoll('a', 'b'), 0)
340340
self.assertEqual(locale.strcoll('a', 'a'), 0)
341341
self.assertGreater(locale.strcoll('b', 'a'), 0)
342+
# embedded null character
343+
self.assertRaises(ValueError, locale.strcoll, 'a\0', 'a')
344+
self.assertRaises(ValueError, locale.strcoll, 'a', 'a\0')
342345

343346
def test_strxfrm(self):
344347
self.assertLess(locale.strxfrm('a'), locale.strxfrm('b'))
348+
# embedded null character
349+
self.assertRaises(ValueError, locale.strxfrm, 'a\0')
345350

346351

347352
class TestEnUSCollation(BaseLocalizedTest, TestCollation):

‎Lib/test/test_time.py‎

Lines changed: 4 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -126,6 +126,10 @@ def test_strftime(self):
126126
except ValueError:
127127
self.fail('conversion specifier: %r failed.' % format)
128128

129+
self.assertRaises(TypeError, time.strftime, b'%S', tt)
130+
# embedded null character
131+
self.assertRaises(ValueError, time.strftime, '%S\0', tt)
132+
129133
def _bounds_checking(self, func):
130134
# Make sure that strftime() checks the bounds of the various parts
131135
# of the time tuple (0 is valid for *all* values).

‎Lib/test/test_winsound.py‎

Lines changed: 2 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -98,6 +98,8 @@ def test_errors(self):
9898
self.assertRaises(TypeError, winsound.PlaySound, "bad",
9999
winsound.SND_MEMORY)
100100
self.assertRaises(TypeError, winsound.PlaySound, 1, 0)
101+
# embedded null character
102+
self.assertRaises(ValueError, winsound.PlaySound, 'bad\0', 0)
101103

102104
def test_keyword_args(self):
103105
safe_PlaySound(flags=winsound.SND_ALIAS, sound="SystemExit")

‎Modules/_ctypes/callproc.c‎

Lines changed: 4 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -1231,14 +1231,15 @@ The handle may be used to locate exported functions in this\n\
12311231
module.\n";
12321232
static PyObject *load_library(PyObject *self, PyObject *args)
12331233
{
1234-
WCHAR *name;
1234+
const WCHAR *name;
12351235
PyObject *nameobj;
12361236
PyObject *ignored;
12371237
HMODULE hMod;
1238-
if (!PyArg_ParseTuple(args, "O|O:LoadLibrary", &nameobj, &ignored))
1238+
1239+
if (!PyArg_ParseTuple(args, "U|O:LoadLibrary", &nameobj, &ignored))
12391240
return NULL;
12401241

1241-
name = PyUnicode_AsUnicode(nameobj);
1242+
name = _PyUnicode_AsUnicode(nameobj);
12421243
if (!name)
12431244
return NULL;
12441245

0 commit comments

Comments
 (0)